Not to be outdone, Google a day earlier announced plans to add a similar AI chat called Bard to its search engine.
Is AI on the verge of changing everything? Yes and no. AI is generating buzz right now for two reasons: First, the latest AI advances can do some eerily human things, like write sonnets, make computer code and create pictures out of thin air. Second, some of this is accessible in products we can all see, and that’s amped up an arms race in Silicon Valley to be seen as the AI leader.
If you haven’t tried these AI products yet, I’ve got a mini guide below to some of the ones worth checking out — and links to my favorite Washington Post stories about how they work.
Still, let’s be clear: Some of what you hear about AI is hype. People have been too quick to attribute human thinking to the current state of the art. The truth is that, for all their abilities, most of these AI products need more work. Case in point: When I searched with Microsoft’s new Bing chat, it sometimes gave me wrong information. I even caught it hallucinating about something that doesn’t exist.
But it’s improving rapidly. The current wave of AI products is built on a technical breakthrough called generative AI. It allows a computer to create images or words that look like they might have been made by a human. It does this by studying zillions of pictures and text samples, often scraped from the web. To use them, you feed them an English-language phrase telling them what you want to do. Figuring out the right prompt is becoming an art of its own.
What’s exciting for all of us is that anyone can do cool stuff with AI now. Generative AI is a creative tool, like the calculator and Photoshop before it. You’re just going to have to learn to speak its language.
Here’s a guide to three areas where you can already get a taste of it. Send me an email to let me know what you’re creating.
It’s made by a company called OpenAI, which has been at the forefront of making generative AI available to the masses.
What it does: A writing tool that responds to queries and requests with text that looks like it was made by a human.
How you get it: It’s free to use on this website, but has been so popular that sometimes it’s unavailable during peak hours. OpenAI plans to charge people $20 a month for a commercial version that’s faster and always available.
What’s the buzz: People are finding all sorts of real-world uses for ChatGPT, from drafting screenplays to writing computer code.
ChatGPT can be easy to talk to, but it’s also bad at math and often strangely confident about being wrong.
Using ChatGPT for anything other than personal experimentation comes with lots of ethical concerns.
You might find ChatGPT useful at work. Help Desk’s Danielle Abril tested me to see whether I could identify emails written by the AI vs. her. (Spoiler alert: I sometimes could not.)
As you might have imagined, students are already using it to cheat.
The product: Microsoft’s new Bing
What it does: Adds a chat interface powered by OpenAI (and similar to ChatGPT) into Microsoft’s search engine and Edge web browser. You can pose complex questions that involve synthesizing information from multiple sources and go back and forth in conversation to refine your question. You can also ask it to get creative and draft emails, poems and more.
How you get it: For now, this version of Bing is only available to people who sign up to join an early-access program here, and also use Microsoft’s Edge web browser on a Windows PC or Mac.
What’s the buzz: I tried Bing chat after Microsoft’s announcement on Tuesday. My takeaway: It can be useful to research complex topics, but its responses are often too wordy to be useful. And there’s a lot of work ahead to make sure it gives answers that are factual and don’t spread misinformation.
And since we’re talking about search, what about Google? The buzz generated by ChatGPT shook the tech giant.
So what does Google have planned with Bard? So far we’ve just gotten a tiny glimpse that looks similar to Bing.
But oops, people already spotted Bard making what appears to be an error in Google’s announcement.
The products: DALL-E 2 from OpenAI and Lensa from Prisma Labs
What they do: Create remarkably authentic-looking drawings, paintings and photographs.
DALL-E 2 turns text prompts into images, such as “painting of koalas in space in the style of Vincent van Gogh.”
Lensa, a phone app, asks you to upload your own photographs to train its AI, and then it generates made-up “magic avatar” images based on them.
All of this is possible because AI can be trained to “understand” huge databases of images and then reinterpret them into new images.
How you get them: DALL-E 2 is available to try here. Users get free credits that refill, and OpenAI sells additional credits.
Lensa is available in Apple’s App Store here free, but it charges to create avatars.
What’s the buzz: Creating images out of little more than a sentence boggles the mind and also opens many cans of worms.
Artists are angry that their work was used to train this AI — and they weren’t compensated.
There’s debate about whether these images should count as “art.”